A system theoretical approach to the analysis of centralized and distributed algorithms for costrained and unconstrained optimisation

ABSTRACT

Popular algorithms for unconstrained and constrained optimization will be presented and analyzed using a control-informed perspective. A system theoretical reinterpretation of the algorithmic schemes as discrete-time dynamical systems will be shown and used to prove the asymptotic convergence of the algorithms to the solution of the specific problem at hand. The extension to a distributed implementation of the algorithms will be also discussed.

SCHEDULE

This course is part of the training provided for IBES PhD Students in the academic year 2023/2024.

Those who succesfully attend and complete the course:

• are eligible for max. 3 credits if they are enrolled in cycle 37;
• gain 8 attendance hours of the IBES training for the Automatic Control and Operational Research (ARO) curriculum if they are enrolled in cycle 38 and 39.

Check the links above for further information about the specific academic requirements for each cyle and how to comply with them.